Scripts

• Freiburg flatten: Flatten the Freiburg dataset (see Datasets > Freiburg)

• Giusti flatten: Flatten the Giusti dataset (see Datasets > Giusti)

• Yamaha flatten: Unused

• Single folder flatten: Used to flatten our own data (like “Steentjes”)

• Remove sequential and similar images: The giusti dataset contains a lot of similar images, this script fixes that

• Convert masks: Convert the multicolor Freiburg and Yamaha masks to our own grayscale masks style

• Images to TFR: Convert images and masks to a single TFRecord file